Abstract—In this research, we analyze nursing care records

by applying an analysis tool KH Coder to nursing care life logs,

in order to develop a method for visualizing and verifying

nursing care actions. We use 161 nursing care life logs

recorded in Long-term Care Health Facility S in M City. We

identify the descriptions related to work contents of nursing

care workers from the nursing care records including various

intentions and contexts. The analysis results using KH Coder

showed that central issues in nursing care were extracted, and

the role of nursing care such as overall structure of various

subjects was clarified. We found that the analysis results have

potential to clarify the work content of care workers. As the

nursing field requires efficiency in health care services,

improvement and continuous data collection are important for

the long-term building of health care services as well as

large-scale data collection. In the future, we aim to develop an

Electronic Medical Record that can be created

semi-automatically in accordance with the level of care

required.

Index Terms— Medical information, Electronic Medical

Record, Text data mining, Nursing care life log, KH Coder.

I. INTRODUCTION

N Electronic Medical Record (EMR) records patient

information by computers instead of by papers. Not

only the data but also the entire management system may be

called EMR. The expected effect simplifies the entire

process of hospital management and improves medical care

[1, 2]. Because data are managed electronically, input data

can be easily managed in comparison with paper-based

medical records [3]. Information can be easily shared

electronically [4, 5]. On the other hand, falsification must be

prevented and the originality of the data must be guaranteed.

Data mining searches for correlations among items by

analyzing a great deal of such accumulated data as sales data

and telephone call histories. Text data mining resembles

data mining because it extracts useful knowledge and

information by analyzing the diversified viewpoints of

written data [6].

Recently, the interest has risen in text data mining because

it uncovers useful knowledge buried in a large amount of

Manuscript received 10. Nov 2018, revised 5. Jan, 2019.

This work was supported by JSPS KAKENHI Grant Number

JP18K11530.

M. Kushima, T. Yamazaki and K. Araki are with the Faculty of Medicine

at the University of Miyazaki Hospital.

e-mail: muneo_kushima@med.miyazaki-u.ac.jp,

e-mail: yama-cp@med.miyazaki-u.ac.jp,

e-mail: taichan@med.miyazaki-u.ac.jp.

http://mit.med.miyazaki-u.ac.jp/.

5200, Kihara, Kiyotake-cho, Miyazaki-shi, Miyazaki 889-1692 Japan.

Tel: +81-985-85-9057, Fax: +81-985-84-2549.

accumulated documents [7, 8].

Fig. 1 Screen shot of an EMR

Research has started to apply text data mining to medicine

and healing [9, 10]. In addition, the speed of electronic

medical treatment data is accelerating because of the rapid

informationization of medical systems, including EMRs.

Recently, research on data mining in medical treatment that

aims for knowledge and pattern extraction from a huge

accumulated database is increasing. However, many

medical documents, including EMRs that describe the

treatment information of patients, are text information.

Moreover, mining such information is complicated. The

data arrangement and retrieval of such text parts become

difficult because they are often described in a free format;

the words, phrases, and expressions are too subjective and

reflect each writer [11].

In the future, the text data mining of documents will be

used for lateral retrieval, even in the medical treatment

world, not only by the numerical values of the inspection

data but also by computerizing documents.

In this research, we analyze the care life logs using an

analysis tool KH Coder for visualizing and verifying nursing

care actions.

II. EMR AT UNIVERSITY OF MIYAZAKI HOSPITAL

Fig. 1 shows a screen shot of an EMR. When the medical

information system was updated on May, 2006, the

University of Miyazaki Hospital introduced a package

version of the EMR system called Integrated Zero-Aborting

NAvigation system for Medical Information, which was

developed in collaboration with a local IT company. The

recorded main data include patient's symptoms, laboratory

results, prescribed medicines, and the tracking of the

changed data. The cases of making both the images of

Text Data Mining of the Nursing Care Life Log

from Electronic Medical Record

Muneo Kushima, Tomoyoshi Yamazaki and Kenji Araki

A

Proceedings of the International MultiConference of Engineers and Computer Scientists 2019 IMECS 2019, March 13-15, 2019, Hong Kong

ISBN: 978-988-14048-5-5 ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online)

IMECS 2019

X-rays and the appended electronic materials are not

infrequent either. If a network is used, EMR can be shared

not only in one hospital but also among two or more

hospitals.

EMR has a unique feature that is different from those being

operated at many other university hospitals.

First, the electronic card systems used so far in university

hospitals were all developed by major medical system

venders, but EMR was developed in collaboration with local

companies. The advantages of collaboration with local

companies included prompt communication and lower

costs.

Second, we focused on performance, especially the speed

at which the screen opens.

Third, we aimed for a useful system to improve

management, reflecting a request by the University of

Miyazaki Hospital after it was incorporated.

We made the medical staff concretely are aware of the cost

and made the management analysis system work closely

with the EMR system and showed its cost when the system

was ordered.

III. TEXT DATAMINING APPLICATION TO MEDICINE

Text data mining is often used to analyze information

hidden in the text of a document and to extract key words,

phrases, and even concepts from written documents. Text

data mining or data mining, which is roughly equivalent to

text analytics, refers to the process of deriving high-quality

information from texts.

Text data mining usually structures the input text (often

by parsing, adding derived linguistic features, removing

others and inserting into a database), derives patterns within

the structured data, and finally evaluates and interprets the

output.

Fig. 2 shows the process of text data mining. Two

particular aspects should be considered when applying text

data mining to a medical context. Second, final decisions

regarding courses of treatment can be obtained.

One difficulty in applying text data mining to medicine is

the entire process of identifying symptoms for

understanding the associated risks while taking appropriate

action.

IV. NURSING CARE LIFE LOG

A Nursing Care Life log records a 24-hour period of the

caregiver’s activity. It is also utilized as a long-term service

content record. The recording itself is not the main purpose,

but it transmits information to others, accumulates and

analyzes data, and aims to lead the service to better care.

The text data of the nursing care record is a text record

integrating the facility service usage record of the care

receiver and the observation record of the care giver. It is

used for the cooperation and transmission to other

occupations and grasp of the state of the care receiver among

the care givers.

Also, due to effective operation and improvement of

nursing care work, education / training of nursing care

workers, development of secondary use of nursing care

records is strongly desired from nursing staff in the field.

In the development of secondary usage of data

accumulated in nursing care records at nursing care facilities,

the amount of text data is enormous, and it was a major

obstacle to organize data systematically. As a method to

overcome this obstacle and to acquire knowledge that can be

used to solve the above problem from enormous text data,

text mining technology has attracted attention.

Fig.2 Process of text data mining

Fig. 3 Example of a screen shot of KH Coder

Generally, a life log is a technique of recording human life,

work, experience as digital data such as video / audio /

position information, or the record itself. In this research, we

use text data recorded at nursing care site.

V. KH CODER

KH Coder is an open source software for computer

assisted qualitative data analysis, particularly quantitative

content analysis and text mining. It can be also used for

computational linguistics. It supports processing and

etymological information of text in several languages, such

as Japanese, English, French, German, Italian, Portuguese

and Spanish. Specifically, it can contribute factual

examination co-event system hub structure, computerized

arranging guide, multidimensional scaling and comparative

calculations.

It is well received by researchers worldwide and used in a

large number of disciplines, including neuroscience,

sociology, psychology, public health, media studies,

education research and computer science.

KH Coder has been reviewed as a user friendly tool "for

identifying themes in large unstructured data sets, such as

Proceedings of the International MultiConference of Engineers and Computer Scientists 2019 IMECS 2019, March 13-15, 2019, Hong Kong

ISBN: 978-988-14048-5-5 ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online)

IMECS 2019

online reviews or open-ended customer feedback" and has

been reviewed in comparison to WordStat.

KH Coder supports various kinds of searches with

frequency tables indicating what kind of words appeared

frequently. Furthermore, the concepts contained in the data

can be investigated by looking at groups of words appearing

together or groups of documents containing the same words,

based on multivariate analysis [12]. Moreover, the

characteristics of the document group can be identified by

listing words which appear particularly frequently in the

document group. It is possible to automatically classify

documents according to criteria designated by analysts. Fig.

3 is an example of a screen shot of KH Coder.

VI. ANALYSIS RESULTS

In order to identify the care worker's work contents and

the related descriptions, we analyze 161 nursing care life

logs including various intentions and contexts, recorded in

geriatric health facility S in M city. In this research, we

evaluate the visualization result to judge whether or not it is

important for the care worker desires.

Text analysis results where the input is nursing care

record text data are shown as below. It is an environment

that can interactively acquire output results according to the

interests of care workers using multiple result diagram

panels.

Table 1 shows frequently occurring words. In Table 1, the

most frequent word was "toilet". Frequent keywords related

to the work content in nursing care facilities were "toilets",

"urination", "doing", "calling", "saying", "morning",

"sleeping", "wheelchair", "voice", "induction", "putting",

"assistance".

Fig. 4 (a)(b) show network diagrams of words and word

connections.

In Fig. 4 (a), "toilet" is a keyword because it is the center

of strong co-occurrence. As for "toilet", the connection of

nursing care was seen mainly from "urination", "induction"

that co-occurred with "toilet".

In Fig. 4 (b), looking at the word arrangement,

"wheelchair", "hole", "night time", and "appearance" were

almost in the center.

In the co-occurrence network, knowledge extraction was

classified into five groups.

From the set of extracted words, group 1 was interpreted

as "toilet", group 2 as "family", group 3 as "assistance",

group 4 as "procedure" and group 5 as "motion".

In Fig. 5, focusing on the toilet in the cluster represented

in the self-organizing map, "toilets", "guidance",

"afternoon", "morning", "participation", "walking",

"recreation", "rest", "pat", "exchange", "hall", "exercise",

"walking", "call", "nurse" were formed in the same cluster.

Fig. 6 shows a diagram in which similar words for

executing cluster analysis are classified in a hierarchical

structure, in order to hierarchically capture combinations of

words having similar appearance patterns from the extracted

words.

Table 1 Frequent occurring words

(a)

(b)

Fig.4 Co-occurrence network

No word frequency No word frequency

1 toilet 21 24 walking 6

2 urination 16 25 visit 6

3 perform 13 26 nurse 5

4 call 12 27 bed 5

5 say 12 28 hole 5

6 a.m. 10 29 exchange 5

7 sleep 10 30 go 5

8 wheelchair 9 31 behavior 5

9 voice 9 32 ED 4

10 induction 8 33 p- 4

11 pat 7 34 diapers 4

12 assistance 7 35 stability 4

13 hand 7 36 push 4

14 night time 7 37 spend 4

15 appearance 7 38 rest 4

16 tube 6 39 afternoon 4

17 recreation 6 40 wait 4

18 transfer 6 41 leg 4

19 watch 6 42 correspondence 4

20 join 6 43 lunch 4

21 body 6 44 injection 4

22 entrance 6 45 doctor 3

23 bathing 6

Proceedings of the International MultiConference of Engineers and Computer Scientists 2019 IMECS 2019, March 13-15, 2019, Hong Kong

ISBN: 978-988-14048-5-5 ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online)

IMECS 2019

Hierarchical cluster analysis was performed with the

minimum number of occurrences limited to 4 or more.

"Induction", "toilet", and "urination" were included in the

same cluster, and other things concerning the elderly home

facilities regarding "wheelchair", "assistance" and "transfer"

were also convincing. This indicates that the numerical

value of the degree of care required is high. In the cluster

analysis, knowledge extraction was classified into six

clusters. From the clustering of knowledge extraction,

cluster 1 was interpreted as "toilet", cluster 2 as "night time",

cluster 3 as "bed", cluster 4 as "wheelchair", cluster 5 as

"sleep", and cluster 6 as "motion".

VII. CONSIDERATION

The following is an overall evaluation.

Text data mining in general or data analysis of EMRs

remains a relatively unexplored field. Greater collaboration

between medical and information sectors will improve the

technology so that it can be applied in clinical practice.

As a result of this research, extracted frequent words are

the theme of this research.

Care records are mainly focused on basic vocabulary in

nursing care. Although it is mere records and memorandums,

it can be shared with other care givers, because it is

described as a general natural language.

It is possible to interpret the state of nursing care by

visualization, and the vocabulary extracted this time is valid

for creating a nursing care dictionary.

By visualizing the relation of the extracted vocabulary, it

shows the possibility of standardizing the nursing care

recording method while characterizing the nursing care

point.

Furthermore, from the present study, it was possible to

suggest a direction to construct an electronic nursing care

system of care records.

VIII. CONCLUSION

In this research, we attempted to examine the extraction of

knowledge that the care worker recognizes, from the care

life logs by using the text mining method.

The analysis results could contribute to the clarification of

knowledge content of a wide range of care workers.

In future work, we further build up research on nursing

care record analysis based on the analysis results clarified in

this research, and build a record database.

ACKNOWLEDGMENT

The authors thank the members of the Medical

Information Department at the University of Miyazaki

Hospital.

Fig. 5 Self-organizing map

Fig. 6 Hierarchical cluster analysis

induction

toilet

urination

a.m.

afternoon

join

recreation

p-

rest

hole

bathing

nighttime

voice

walking

leg

pat

exchange

entrance

spend

push

perform

call

nurse

bed

watch

diapers

correspondence

wheelchair

assistance

transfer

stability

have

say

appearance

visit

hand

body

sleep

tube

ED

lunch

motion

injection

go

Proceedings of the International MultiConference of Engineers and Computer Scientists 2019 IMECS 2019, March 13-15, 2019, Hong Kong

ISBN: 978-988-14048-5-5 ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online)

IMECS 2019

REFERENCES

[1] Information System,” Japan Association for medical Informatics, vol.

22, no. 4, pp. 347-353, 2002.

[2] K. Yamamoto, S. Matsumoto, H. Matsuda, H. Tada, A. Matsuyama,

K. Yanagihara, S. Teramukai and M. Fukushima, “Development and

Prospects of Data Capture System for Clinical Study by the Secondary

Use of Electronic Medical Records,” Japan Association for medical

Informatics, vol. 27, no. 2, pp. 211-218, 2007.

[3] Y. Matsumura, H. Nakano, H. Kusuoka, K. Park, M. Matsuoka, H.

Oshima, M. Hayakawa and H. Takeda, “Clinic Hospital Cooperation

System Based on The Network Type Electronic Patient Record,”

Japan Association for medical Informatics, vol. 22, no. 1, pp. 19-26,

2002.

[4] S. Murayama, K. Okuhara and H. Ishii, “Innovation in Manufacturing

Premise by New Finding Obtained from Accident Relapse Prevention

Report,” Proceedings of The 13th Asia Pacific Management

Conference, Melbourne, Australia, vol. 13, pp. 1124-1129, 2007.

[5] Y. Takahashi, K. Miyaki, T. Shimbo and T. Nakayama, “Text-mining

with Network Analysis of News About Asbestos Panic,” Japan

Association for medical Informatics, vol. 27, no. 1, pp. 83-89, 2007.

[6] M. Usui and Y. Ohsawa, “Chance Discovery for Decision Consensus

in Company - Touchable, Visible, and Sharable for a Textile

Manufacture Company -,” Japan Society for Fuzzy Theory and

Intelligent Informatics, vol. 15, no. 3, pp. 275-285, 2004.

[7] H. Abe, S. Hirano and S. Tsumoto, “Mining a Clinical Course

Knowledge from Summary Text,” The 19th Annual Conference of the

Japanese Society for Artificial Intelligence, vol. 1F4-04, pp. 1-2,

2005.

[8] Y. Kinosada, T. Umemoto, A. Inokuchi, K. Takeda and N. Inaoka,

“Challenge to Quantitative Analysis for Clinical Processes by Using

Mining Technology,” Data Science Journal, vol. 26, no. 3, pp.

191-199, 2006.

[9] H. Ono, K. Takabayashi, T. Suzuki, H. Yokoi, A. Imiya and Y.

Satomura, “Classification of Discharge Summaries by Text Mining,”

Japan Association for medical Informatics, vol. 24, no. 1, pp. 35-44,

2004.

[10] Y. Sato, H. Takeuchi K. Hoshi, N. Uramoto, T. Satoh, N. Inaoka, K.

Takeda and N. Yamaguchi, “The Effectiveness of the Text Mining

and Similar Document Search System for Evidence-Based Guideline

Development,” Japan Association for medical Informatics, vol. 24,

no. 2, pp. 315-322, 2004.

[11] M. Kushima, K. Araki, M. Suzuki, S. Araki and T. Nikama, “Analysis

and Visualization of In-patients' Nursing Record Using Text Mining

Technique,” International MultiConference of Engineers and

Computer Scientists 2011 (IMECS2011), vol.2188, pp. 436-441,

Hong Kong, March, 2011.

[12] Tsukada, S. and Morita, T. 2018 "A Study of Public Opinion on Green

Spaces Using Data from Free-Text Descriptive Responses- a Case

Study of Regional City in Japan" International Journal of GEOMATE

14: 21–26.

Proceedings of the International MultiConference of Engineers and Computer Scientists 2019 IMECS 2019, March 13-15, 2019, Hong Kong

ISBN: 978-988-14048-5-5 ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online)

IMECS 2019